Construction and validation of the prognostic model for patients with gallbladder neuroendocrine neoplasm-based a nationwide retrospective registry study.

Abstract

e16140 Background: Gallbladder neuroendocrine neoplasm (GB-NEN) is an extremely rare type of gallbladder neoplasm and the prognosis of patients with GB-NENs varies widely. The focus of this study was to construct and validate a prognostic model of GB-NENs based on the data from a nationwide retrospective multicenter registry cohort. Methods: The Chinese Research Group of Gallbladder Cancer (CRGGC) study is a multicenter retrospective registry cohort study reviewing the electronic medical records from 76 tertiary and secondary hospitals across 28 provinces in China. We assembled patients with GB-NEN diagnosed between 2010 and 2017 in 24 tertiary hospitals from CRGGC study. Influential factors of recurrence free survival (RFS) and overall survival (OS) were identified using Least absolute shrinkage and selection operator (LASSO) regression and Cox predicting model was constructed for predicting RFS and OS. The prognostic model was validated both internally and externally. Performance of the model was assessed by concordance index (C-index) and calibration curves stratified by risk categories. Results: A total of 84 patients with GB-NEN was extracted from CRGGC study, among which 3 were diagnosed with GB-NEN Grade 1, 8 were diagnosed with GB-NEN Grade 2 and 35 with GB-NEN Grade 3. Of the 35 patients with GB-NEN Grade 3, 3 had poorly differentiated small-cell neuroendocrine carcinomas (SC-NECs) and 8 patients had Mixed neuroendocrine–non‐neuroendocrine neoplasms (MiNENs). 53 patients were included in the training group (median survival, 21.1 months; 95% confidence interval (CI), 13.8-37.6 months) and 31 patients in the validation cohort (median survival, 34.3 months; 95% CI, 9.63-58.97months). A Cox predicting model was built, including 5 variables (age, TNM stage, surgery, metastasis at liver, proportion of positive lymph nodes examined). The C-index was 78.1 (95% CI, 63.9-92.3), with a good calibration, and these results were confirmed in the internal validation cohorts. Conclusions: We have built a clinical model for GB-NEN patients and clinical doctors for prognosis prediction based on age, TNM stage, surgery, metastasis at liver, proportion of positive lymph nodes examined which can be easily obtained in one’s case history. It can be used to determine subgroups of patients who may have poorer prognosis and assist individualized and precise treatment for GB-NEN.